Technical-ish question/request about Book of Star Mythology scoring

[who_is_that]

Hi, I made an account to ask this, since I figured this would be a good place. I'm not really sure if this counts as tech support, but please move it there if it does.

I have recently been putting a lot of time into scoring the Extra stage of Book of Star Mythology. After getting a score of 14.7 billion, I downloaded the replay file for the 15.1b world record by Lidera with the intention to study it. You can view the author's ending screenshot and download the replay from their Twitter post about it: https://twitter.com/lidera_bw/status/814713829766483968

But, for some reason, even though this replay file indeed displays a score of 15.1b, it plays out very differently - it scores only 14.3b, ends the stage with exactly 10,000 less PIV than in their Twitter screenshot (229,280 -> 219,280), with the magnitude multiplier (x7.20) and graze (6,836) being identical. These are all lower numbers than in my 14.7b replay, which ended with 221,810 PIV, a x7.25 magnitude multiplier, and 6,853 graze.

https://drive.google.com/drive/folders/1bKbZq2TSJkTK_thta7cTLSwHv6kYe1k-?usp=sharing
Here is a Google Drive folder with videos of Lidera's replay and my replay as recorded on my computer, along with the .dat replay files I used for both runs. His is 09 and mine is 93. I used individual clean copies of BoSM that were downloaded from the Axfc link from the English wiki page. BoSM replays break if you don't use the right "effect" setting, so if you play these back yourself, go into the options menu and change effect to 0 before watching Lidera's replay and 4 before watching mine. Also, put them in a /script/th_bsm/replay folder if you don't have one.

This doesn't seem to be a matter of desyncing, since both runs clear and there aren't any wacky deaths or anything. Though, Lidera not shooting down the remaining rocks at the end of the timeout phase 6 is a bit strange. Additionally, if you pit our replays against each other, his score being higher doesn't seem to make sense. He enters the midboss with more score and 2,000 more PIV, and exits it with 100m more score than me. However, I quickly catch up during the boss thanks to my higher midboss exit PIV for more SCB combined with better shotgunning for more wish items. We both enter the timeout spell with nearly identical score, PIV, and graze, and I have a superior performance from there on, getting more wish items in every final chapter, and getting slightly more SCB and wish items during the last word.

I'm just really frustrated about this. I opened this replay ready to see some amazing things, only to see gameplay that's generally not as polished as my own other than the much better early game routes, except he magically gets (exactly!) 10k PIV from literally nowhere. Now I feel totally lost on how to beat this 15.1b score. Is something wrong with my game? Was something wrong with Lidera's game? Is there something I'm overlooking? I wish I knew...

Thank you to anyone who might be able to help with this.

[williewillus]

What version of the game are you running?

Replays are sensitive to the version of the game they were played on, so there might be differences there.

Also, I remember hearing that some clownily-coded danmakufu games desync based on the effect cut settings, so I would try to match the settings in his video to the best of your ability.

[who_is_that]

Thank you for the quick reply!

What version of the game are you running?

I did my run and played back both replays on v1.10. As far as I know, this is the only version of the game that has the extra stage, so I assume it's the one he used, too.

Also, I remember hearing that some clownily-coded danmakufu games desync based on the effect cut settings, so I would try to match the settings in his video to the best of your ability.

Yep, this happens in the game. His replay desyncs starting at Fuma spell 3 with effect cut set to anything but 0, and mine desyncs at the same place with effect cut set to anything but 4. Both seem to work without issue under the correct settings.

I suspected this might have something to do with low effect cut not properly creating bullet cancel PIV items or something, as that's the only source of PIV that isn't quantified by the HUD, and thus could unknowingly disappear. But I don't think this is the case, as Lidera's replay enters Fuma spell 3 with 103,940 PIV regardless of the effect setting, as done mine with 104,460.

There's also a background setting with three options (All, Simple, None) that I tried in combination with 0 effect on Lidera's replay, but there weren't any score, PIV, or gameplay differences. Though I didn't try every possible combination of effect cut settings and background settings on both replays. The only other in-game options are Rebirth: Home/Stay (which only affects the main mode's pointdevice mode), a setting for which key unlocks the framerate (which I highly doubt affects anything), and an option to change the game speed. I left the one untouched and assume Lidera left it untouched in the way PCB's slow mode is. But I went ahead and ran through Lidera's replay a couple times with 0 effect testing these settings, and none of them turned out any differently.

[degradingwindow]

I checked both replay files on clean copies as well and I have not noted any differences from what you've described.

Although I don't know the exact inner workings of the ph3 replay files (something I'll have to look into later), what I do know is that things that don't affect the main gameplay such as backgrounds won't affect your replays. Game speed setting also won't apparently mess up the replay as the replay file automatically stores the setting (I really do have to look into it). I'll try to see if the replays are different in the original Japanese version but I highly doubt it would change anything.

I don't have the technical knowledge to crack open the replay files and tell you how Lidera got such a high score in the latter half of the game but if it helps, I could try to take a look at the scoring system in the game and tell you how it exactly works down to the minute detail.

[who_is_that]

Thank you for taking a look at the replays! I was hoping to get people to check them in hopes that somebody's setup might play it back differently. But between all my own efforts and the people I've asked, including this thread, it doesn't seem like that'll lead to much. I'm probably just going to buckle down and find my own improvements (most likely graze milking)

I'll try to see if the replays are different in the original Japanese version but I highly doubt it would change anything.

I've done all my playbacks on the unpatched original Japanese version, so I also doubt anything would change.

if it helps, I could try to take a look at the scoring system in the game and tell you how it exactly works down to the minute detail.

If you want to, I would appreciate that! I'm not sure how much you could uncover, since if there was some obscure mechanic that was used to gain PIV, surely it would show up in the replay... but please feel free to post anything you find.